Response to Reviews: A sampler for atmospheric volatile organic compounds

by copter unmanned aerial vehicles

Karena A. McKinney, Daniel Wang, Jianhuai Ye, Jean-Baptiste de Fouchier, Patricia C.

Guimarães, Carla E. Batista, Rodrigo A. F. Souza, Eliane G. Alves, Dasa Gu, Alex B. Guenther,

and Scot T. Martin

Reviewer comments are in bold, the authors’ response is in Plain text, and revisions to the

manuscript are in Blue italics. Please note that for consistency, all line numbers refer to the

version of the manuscript published in the Discussion forum.

Reviewer 1

Overall, this is a well-written paper and a valuable technology. It should be published with

minor revisions, however, there are some important discussion points and details that I

would like to see addressed.

We thank the reviewer for the extremely helpful and detailed comments, which have led to

significant improvements to the manuscript. Our responses to the comments and corresponding

changes to the manuscript are described below.

General Comments: 1) The dilution due to rotor-wash, which is a problem for all

instruments without an inlet that extends beyond the turbulence induced by the multi-rotor

platform, is not discussed until later in the paper.

The issue of influence of rotor-induced turbulence and the need for CFD simulations to

understand its effects have been introduced earlier in the paper. Specifically, we have made the

following additions to the Abstract and Introduction:

Abstract: “The species identified, their concentrations, their uncertainties, and the possible

effects of the UAV platform on the results are presented and discussed in the context of the

sampler design and capabilities.”

Introduction, line 71: “As with any new sampling method, the possible introduction of artifacts

due to the platform should be considered. For the case of UAVs, as with manned aircraft, the

platform itself disturbs the surrounding air, which could lead to issues such as loss of target

species on surfaces, outgassing of interfering species, or artifacts in measured concentrations

due to enhanced mixing of the sample air.”

Introduction, line 100: “The possible effects of the UAV platform on the surrounding air and

thereby on the collected sample are an important consideration which is explored by

computational fluid dynamics simulations.”

The authors conclude that their samples are representative of ambient mixing ratios;

however, while this may be the case for isoprene and monoterpenes, the carbon fiber DJI

M600 Pro body likely has some emissions of low molecular weight VOCs, which could pose

problems for cartridge measurements of other VOCs.

Any emissions from the carbon fiber of the drone would be diluted in the ambient air flow past

the drone, which is large due to the motion of the rotors. We therefore expect that any resulting

interferences would be small. It is nonetheless worth investigating. One way of assessing such

artifacts would be the blank cartridges obtained during flight. We have added to the Supplement

a table of VOC masses detected in the cartridge samples and blanks collected onboard the UAV

and from the tower. A comparison between the blanks obtained in flight and on the tower shows

similar background levels, suggesting that outgassing from the drone does not interfere with the

targeted VOCs. We have added the following statement to Section 2.4:

Line 239: “A comparison of the chromatograms of samples and blanks collected by the sampler

with those collected on the tower does not indicate the presence of any artifacts in the sampler

cartridges attributed to outgassing of volatile compounds from the drone.”

Second, simulations are shown for the legs extending in the landing position, although I

imagine samples were collected when the legs are retracted (as is done automatically by the

M600 software after takeoff). The differences in the flow with legs retracted or if samples

were collected when the legs were in the landing position should be discussed.

In order to put it in the context of other changes, we respond to this comment in more detail

below. Please refer to the comment on P22 (Figure 4).

2) The challenges associated with desorption of VOCs and OVOCs from cartridges and

quantitative measurements of these compounds compared with whole air samples should

be discussed.

There are numerous other studies of cartridge sampling that have addressed this issue. We

consider it outside the scope of this study to include a full review here. Instead, we have added a

few sentences to the text referring the reader to some of the key publications on this topic and

comparing the pros and cons of adsorbent cartridges vs whole air samples.

“Woolfenden (2010b, a) and Pankow (2012) review the performance of adsorbent cartridges for

quantitative VOC measurements and compare their retention and recovery of VOCs with whole

air samples. Although whole air canisters have the advantage of a very short (second) fill time,

they are large (1 L volume) and heavy. Adsorbent cartridge samples require longer sampling

times, but their small size and light weight make them well suited to carrying on a UAV.”

3) Please discuss how atmospheric temperature was measured. For instance, what sensor

was used to measure temperature, and was this done in the flow path as well or elsewhere

on the sampling platform? This appears to be a critical measurement for determining the

mixing ratios of VOCs, and it is not explicitly described anywhere.

We thank the reviewer for this comment, which drew our attention to an aspect of the

measurement description that was unclear. Temperature was not measured by the sampler, but

because the flow sensor measures mass flow, not volume flow, temperature is not needed to

calculate moles of sampled air or VOC mixing ratio (nor is pressure). Mass flow sensors operate

by measuring the dissipation of heat by the gas flow. This quantity depends on the mass (or

moles) of gas passing the sensor element per unit time, so it inherently accounts for changes in

temperature and pressure. Mass flow is reported in standard cm3 min-1 (sccm), which can be

converted to moles of gas per minute using the molar volume of an ideal gas at standard

conditions (273 K and 1 atm). Hence, measured ambient temperature and pressure are not used

in the calculation. We have revised the text to remove references to calculating the volume flow

rate and to clarify how the mixing ratio calculation was performed. The revisions are shown

below.

Line 149: “The mass flow rate is converted into a volumetric flow rate using the measured

pressure at the flow sensor and atmospheric temperature. The sample volume is obtained by

integrating the volumetric flow rate over time. The mass flow sensor is used to calculate the total

moles of gas in each sample (c.f., Section 2.4).”

Line 154: “The measured pressure is also used with atmospheric temperature to convert mass

flow rate to volumetric flow rate as UAV altitude changes used as a diagnostic of proper

operation of the flow system.”

Line 228: “The mixing ratio XVOC of VOCs is calculated from the measured mass of each

compound in the sample and the volumetric flow rate according to the following governing

equation:

XVOC = moles VOC / moles air = (mVOC R T) / (MVOC P Q τ) (Eq. 1)

where mVOC is the mass of the VOC measured in the sample, MVOC is the molar mass, R is the gas

constant, T is the temperature, P is the pressure, Q is the volumetric flow rate, and τ is the

sampling time. The mass flow sensor reports the equivalent volume of gas flow per unit time at

standard temperature and pressure conditions (273 K and 1 atm). Inserting these constant values

in Eq. 1 and combining them with R gives:

XVOC = moles VOC / moles air = (mVOC × 22400 sccm/mol) / (MVOC Qstd τ) (Eq. 2)

where Qstd specifies mass flow. Thus, the measured quantities used in calculating XVOC are the

mass of VOC in the sample mVOC, the mass flow rate Qstd, and the sampling time τ. In practice,

since the mass flow rate can vary over the sampling period (Figure 3), a time integral of the

measured mass flow rate is used.”

4) A comparison of samples and blanks would be very useful in demonstrating the utility of

this platform.

A table of measured VOC masses in the samples and blanks has been added to the Supplement

(Table S2) and has been referenced in the text. The table shows that for isoprene and α- and β-

pinene, the mass of VOC in the samples is well in excess of that in the blanks. We have also

updated the data in both Table 1 and Table S2 based on a re-analysis of the original GC data. In

doing so, we noted that in the original manuscript the mixing ratio values in Table 1 were based

on a preliminary analysis rather than the final calibration data. As a result of applying the final

calibration data, the mixing ratios have changed substantially, in some cases by a factor of 2. The

revised values are accurate to within the stated uncertainties.

Specific Comments:

Line 2: Word Choice. Why “copter technology,” not “multi-rotor”?

We have taken the reviewer’s suggestion, but have changed the wording to “multicopter” rather

than “multi-rotor” as multicopter is consistent with the terminology that is used in the title and

throughout the manuscript.

Line 10: The phrase “close to 2 ppt” is vague. Please be more specific, and include the

“3ppt or 20% (total) uncertainty in measured mixing ratios” in the abstract.

The sentence in the abstract has been revised to read: “The overall minimum detection limit for

the sampling volumes and the analytical method was 3 ppt and the uncertainty the greater of 3

ppt or 20% for isoprene and monoterpenes.”

Line 27: delete “and” and insert comma and “from” before “tethered balloons”

The suggested revision has been made.

Line 30-31: Which is less well characterized, horizontal gradients or vertical gradients at

these scales? Discuss which of these is more important for models.

Neither is particularly well characterized. Most measurements are made from towers, so most

represent a single point both horizontally and vertically (some tall towers have multiple points in

the vertical). A single tower observation is often assumed to be representative of a large

geographical area or land cover type. The extent to which this is true has not been fully

investigated and can depend on the region, with the tropics exhibiting greater horizontal

heterogeneity than temperate forests. Emission models are 2-dimensional (i.e., land surface

only). The most widely used of these, MEGAN, has a horizontal resolution of 1 km. The

resolution is based on land cover data and emissions are calculated based on the distribution of

plant types at each grid point. (Emissions are not directly interpolated from tower measurements,

though these measurements can be used to validate the model.) Thus, it would likely be

straightforward to use measurements with higher horizontal resolution to test and improve

existing emission models, which would be an important advance. On the other hand, regional or

global models do not resolve near-canopy vertical gradients in VOCs. This mainly done only in a

small number of (generally 1-D) canopy-scale models that have been used in isolated studies.

Vertical gradients may therefore not impact models as directly, but the results are important for

understanding the interplay of mixing, deposition, and chemical processes in determining the fate

of VOCs, and therefore for informing model development more generally. We have added

comments on this topic to the text. They are shown combined with revisions in response to the

next comment, below.

Line 31-35: “Thus, this scale . . .global atmosphere” Pease rewrite these sentences, as they

read awkwardly. Also, what does “the primary scale for VOC emission” mean? Is that the

finest resolution that models are able to represent? Also, “precisely the missing link”

maybe be slightly overstating the importance of these measurements to understanding of

VOCs in atmospheric chemistry (i.e. we don’t know if this is the “only” missing link, and

indeed, it likely is not). Finally, if these measurements are scarcer in the amazon then

elsewhere, cite some studies that have adequately captured this horizontal or vertical

resolution in other parts of the world, and discuss how it has informed our understanding

of regional emissions and the subsequent atmospheric chemistry.

As suggested by the reviewer, we have rewritten these sentences. Individual VOC measurement

sites are scarcer in the Amazon than elsewhere, but we know of no existing data sets anywhere

that capture the horizontal heterogeneity of forest emissions with a resolution of 10’s to 100’s of

meters. We have clarified this point in the revised text:

“As such, detailed information on the spatial distribution of emissions at 10’s to 100’s of meters

has been difficult to obtain. This information is most critically needed in globally important and

highly spatially heterogeneous source regions of VOCs, such as the Amazon, which is not well

characterized even at large spatial scales. Thus, this scale is not represented in current VOC

data sets, yet it reflects the primary scale is critical for understanding and quantitatively

modeling VOC emission and uptake and is precisely the missing link in vital to advancing our

present-day understanding of VOCs in atmospheric chemistry. This information is even more

scarce in remote areas, such as the Amazon rainforest, that are very important sources of VOCs

to the global atmosphere. New VOC measurements with increased horizontal coverage and

resolution that could be used to test and improve existing emission models would be extremely

valuable. In addition Similarly, knowledge of VOC concentrations as a function of altitude

height throughout the boundary layer over a range of underlying land cover types is needed to

better constrain emissions, chemical reactions, and atmospheric mixing of these compounds and

to thereby inform atmospheric chemistry model development.”

Line 35: replace “height” with “altitude”

The suggested revision has been made.

Line 55-77: Although there are a number of advantages to multirotor UAV platforms, it

would be helpful to discuss the importance of rotor-wash and potential of sample dilution

due to rotor-wash (see general comment 1). I see this is in part addressed later in the paper,

however, this should also be mentioned in the introduction.

As discussed in response to General Comment 1 above, we have added language addressing the

importance of rotor-induced mixing for sampling and motivating the CFD simulations in the

introduction. We have also expanded the discussion of the implications of the CFD simulations

in the Results and Discussion section. These changes are described in more detail below, in

response to the comment on Line 262.

Line 92-92: Is the detection limit of the VOCs entirely determined by the subsequent

analysis (e.g. GC-MS or GC-ToF-MS)?

No, the detection limit is determined by the detection limit of the analysis method, combined

with the background levels measured for the field blanks. The uncertainty in the measurement

also depends on the uncertainty in the measured flow rate. These factors are detailed in Section

2.4. This section of the Introduction is intended as a demonstration that drone-based cartridge

sampling is feasible, not as a detailed discussion of uncertainties. We have, however, added a

sentence regarding the role of the measured VOC background level in determining the detection

limit to the text, as follows:

“The primary scientific requirement of the sampler is that the total mass of analyte collected be

greater than the detection limit of the analytical system for that compound. In the case of a

volatile organic compounds detected by GC-MS, the detection limit has typically been ca. 10 pg.

For a sample volume of a few liters of air, which can be collected in 5 to 15 min by typical flow

rates through adsorbent cartridges, this corresponds to a VOC detection limit of less than 10

pptv (Pankow et al., 2012). Commercial detectors are now available with detection limits of < 1

pg, including the GC-ToF-MS used for this study (Hoker et al., 2015), implying an order of

magnitude lower detectable VOC mixing ratios. The method detection limit also depends on the

background level of VOC measured in field blanks, which is also ca. 10 pg VOC. This

corresponds to a VOC detection limit of less than 10 pptv for a sample volume of a few liters of

air, which can be collected in 5 to 15 min by typical flow rates through adsorbent cartridges

(Pankow et al., 2012).”

Line 93-94: This sentence isn’t needed and is vague (please delete): “this suggests that

detection of VOCs from multicoptor flight. . .”

This sentence is the conclusion of the preceding exercise in determining the required sampling

time and demonstrating that drone-based adsorbent cartridge sampling is feasible. We think it is

an important point, so we have chosen to retain it. To reduce vagueness, we have specified the

drone flight duration to which we are referring. The revised text reads as follows:

“This suggests that detection of VOCs in cartridge samples collected within current multicopter

flight durations of ca. 30 min is feasible.”

Line 95: insert “cartridge” prior to “sampler”

The suggested revision has been made.

Line 115 and Figure 2: Label and discuss the 18V supply from the DJI M600 pro to the

cartridge sampler, and its integration.

The figure has been revised to more clearly label the 18V power supply from the UAV. The

relevant section of the figure caption has been revised as follows:

“All components are powered by onboard batteries on the UAV batteries through the 18 VDC

power output on the Matrice 600 and are controlled by an Arduino Uno microcontroller.”

The section of the text referenced here (Line 115) is intended to be a description of the drone

platform itself. The electrical interface to the sampler is discussed in a later part of Section 2.4

labeled Electrical System. To address the reviewer’s comment, we have revised that section to

more clearly describe the electrical interface between the drone and the sampler. Please see the

response to the comment on Line 172 below for the revised text.

Line 139: Delete “the” before “cartridge sampling”

The suggested revision has been made.

Line 152: Please comment in the text (here) on whether in the future, the use of filters prior

to the cartridges could be helpful in preventing debris from making its way into the system.

I see, filters are finally mentioned on Line 195, however, I think this should be discussed

more fully and earlier.

There is a statement in the text that the flow sensor can be used as an indicator of a malfunction

such as blockage of the flow by debris. In practice, this issue has not arisen during use of the

sampler. We have elected not to use a filter on the inlet since filters can adsorb and later desorb

semi-volatile VOCs, leading to artifacts. On balance, the disadvantage of potential filter artifacts

outweighs the benefit of using a filter to prevent the low-probability chance of obstruction. We

have added several sentences, shown below, explaining this reasoning to the existing text. The

discussion of filters appears in Section 2.3 Sampling Methods. After careful consideration, we

have elected not to move this material earlier in the manuscript. Line 152 is in Section 2.2, which

is a description of the sampler design and operation. The use of filters is more germane to the

discussion of sample handling in Section 2.3. Moving it earlier would, we believe, lead to less

clarity between these two topics.

“No particle or ozone filter was used upstream of the cartridges to prevent loss of analytes on

the filter surfaces. Although an inlet filter could be useful in preventing debris from entering the

sampling system, filters can also adsorb and later desorb semi-volatile VOCs, possibly

introducing sampling artifacts (Zhao et al., 2013). As this was judged to be a greater drawback,

an inlet filter was omitted. As such, both gas- and aerosol-phase VOCs are sampled; the

reported concentrations represent the sum of these contributions. The presence of ozone in the

sample cartridges may contribute to oxidation of the most reactive VOCs between collection and

analysis. The use of an ozone filter may help to mitigate this effect. The effect of ozone filters on

the samples is therefore being evaluated in ongoing work.”

Line 156: Please comment here on how atmospheric temperature was measured (see

general comment 3)?

Please see the response to general comment 3 above.

Line 157: “It outputs analog voltage. . .” Is the same is true of the mass flow sensor, as well

(i.e. produce an analog voltage that is converted into a flow value? Also, is this conversion

based on laboratory or manufacturer based calibrations? Please comment in the text.

Yes, the mass flow sensor also outputs an analog voltage. We have revised the text to clarify this.

The conversion of the mass flow sensor is based on periodic laboratory calibrations. A sentence

to this effect was included in the original version (see below). The revised text reads as follows:

“A mass flow sensor (Model D6F-P; Omron) was installed upstream of the pump to provide a

continuous analog voltage output signal corresponding to the mass flow at standard temperature

and pressure. The flow sensor supports a flow range of 0 to 1000 sccm and includes a built-in

cyclone dust segregation system, which diverts particulates from the sensor element. The mass

flow sensor was calibrated periodically against a reference standard in the lab.”

Because the flow sensor measures mass flow, not volume flow, the data from the pressure sensor

is not used in the VOC mixing ratio calculation. We have clarified this in our response to

General Comment 3, above, and in the revised text. The pressure sensor is therefore used only

for diagnostic purposes (i.e., to determine whether the flow system is functioning properly).

Hence, the factory calibration was deemed sufficient for conversion of the pressure sensor signal.

The description of the pressure sensor was modified to reflect this:

“Pressure system. An absolute pressure transducer (MX4100AP; NXP) is positioned adjacent to

the flow sensor in order to measure the pressure in the flow path. The measured pressure is used

also used with atmospheric temperature to convert mass flow rate to volumetric flow rate as

UAV altitude changes.as a diagnostic of proper operation of the flow system. The device

operates across a pressure range of 20 to 105 kPa. It outputs an analog voltage signal recorded

by the microcontroller that can be converted to a pressure value using a function provided by the

manufacturer. Laboratory calibration of the pressure sensor is possible but was deemed

unnecessary due to its purely diagnostic function.”

Line 162: Please comment on the inline, wetted of solenoid valves and their potential VOC

emissions to which cartridge samples could be exposed. Could this influence the detection

limit of this system, particularly with sensitive analyzers such as GC-ToF-MS?

The solenoid valves and all other wetted parts of the sampling system are positioned downstream

of the sorbent cartridges so that the sampled air does not contact any sampler surfaces prior to

passing through the cartridge. Hence, any contamination due to the solenoid valves or other flow

system materials would only occur diffusively, and would also appear in the field blanks. We

have not observed any such signals in the blanks that have interfered with detection of the target

compounds or affected the detection limit beyond the levels already noted for the blanks (ca. 10

pg, ca. 2.5 pptv). To address this comment in the manuscript, we have added a statement at the

beginning of Section 2.2 Sampler Description stating that the sorbent cartridges are positioned at

the input of the sampler flow path to minimize contamination:

“The adsorbent cartridges are positioned at the inlet of the flow path to ensure that the sample

air does not come in contact with any flow path surfaces prior to sampling as it could lead to

contamination or loss of analytes.”

We have also slightly modified the text in Section 2.3 Sampling Methods where this issue is

discussed. The modified text is as follows:

“The sorbent cartridges are mounted at the sampler inlet to ensure that the sample gas that

passes through the cartridges has not contacted other surfaces in the flow system, thus

preventing potential analyte losses or contamination from the flow system tubing components.”

Line 170: Are there additional sensors to system pressure and system flow on the sampling

platform? If not, please specifically list these two sensors.

No, the pressure and flow sensors are the only two. The suggested revision has been made.

Line 172: “via the power distribution board” is awkward phrasing- consider rewording.

As suggested, we have revised the text as follows:

“The sampling system is powered by the UAV batteries via the 18 VDC power output on the

Matrice 600. The UAV power supplies two voltage regulators which provide 5 VDC output for

the pump, pressure and flow sensors, and Arduino Uno, and valve driver boards, and 24 VDC

output for the valve manifold.”

L182- L189: Discuss the benefits of be able to measure high molecular weight compounds

(C9-C30) of this approach, compared with others.

Two major classes of biogenic VOCs, monoterpenes (C10) and sesquiterpenes (C15), fall in the

C9-C30 range. Hence, the higher molecular weight range is needed to capture these and other

potential compounds of interest. The text has been amended as follows to make this point more

clearly:

“Tenax TA is a relatively weak sorbent that collects components with volatility less than benzene

(e.g., >C6) including monoterpenes, C10, and sesquiterpenes, C15, whereas Carbograph 5TD

shows strong sorbate affinity and captures low-molecular-weight VOCs with carbon number of

C3 to C8 (Woolfenden, 2010a) including isoprene, C5. The combination of these sorbent

materials enables sampling of VOCs with carbon number from C3 to C30, covering the expected

range of atmospheric compounds from biogenic and anthropogenic sources (Goldstein and

Galbally, 2007).”

Line 204-207: Do you base your sample volume collection on prior measurements in

different environments? Can this be adjusted easily in the field or between flights?

The sample volume is determined from the detection limit of the adsorbent cartridges based on

past studies and on the desired detection limit for the VOC mixing ratio. This is discussed in the

introduction on lines 86-91 and as applied to determination of the sample volume for this study

on lines 196-206. Both the flow rate and the sample time affect the sample volume; both are

easily adjustable in the field. We have made the following revisions to the text to clarify these

points:

Lines 141-143: The volumetric flow of the pump is a function of the pressure drop across the

inlet and outlet, and is controlled via a manually adjustable pinch valve (Model 44560; US

Plastic Corp.) at the output of the flow system.

Lines 196-206: The total sample volume depends upon the flow rate and sample collection time.

Both of these parameters are easily adjusted in the field between flights. The flow is adjusted

using the manual pinch valve downstream of the pump. The flight time is programmed in the

flight algorithm executed by the Arduino Uno microcontroller. A constant low volumetric flow

rate is required to allow for optimal sorbent-sorbate interaction and uptake onto the sorbent

matrix. A target flow rate of 150 sccm was defined to maximize both VOC capture efficiency and

sample volume (Woolfenden, 2010b;Markes International Ltd., 2014). Based on the relationship

between sample volume and minimum detection limit reported by past studies (ca. 10 pg,

Pankow et al., 2012), a minimum sampling volume of 1.5 L per adsorbent cartridge collected,

corresponding to ca. 2.5 ppt VOC, is targeted. This results in 10 min of sampling time per

cartridge. Two to three cartridge samples of this volume can be collected in a single flight while

also carrying out take-off/landing and transits between sampling locations. The Arduino Uno

microcontroller provides the operational flexibility to obtain smaller or larger sample volumes

by utilizing either more tubes and shorter collection times or fewer tubes and longer collection

times, respectively, during a single flight.

Line 213-215: “not influence the results”- can you expand on this?

This references a sentence regarding the introduction of sample artifacts due to transport and

storage. The study protocol followed established methods that have been shown to have minimal

artifacts due these factors, and we expect the same to be true in this case. After consideration,

however, we have removed the phrase referenced by the reviewer, which cannot be proven. The

sentence now reads:

“Under proper transport and storage, sample artifacts were have been shown to be minimal and

did not influence the results (Pollmann et al., 2005).”

Line 221-222: Are these internal standards injected prior to sample collection as well or

simply prior to sample analysis? Please explain this in the text.

The internal standards are injected prior to sample analysis. To clarify, the text has been

amended as follows:

“Internal standards tetramethylethylene and decahydronaphtalene are injected into each sample

after collection and prior to analysis.”

Line 240: This is a good description of the uncertainty and the detection limit. This

detection limit and uncertainty do not seem compatible with the “nearly 2 ppt” listed in the

abstract. Are they? If so, please explain.

As noted above, the Abstract has been revised to reflect the 3 ppt detection limit, so it is now

consistent with the values presented here.

Line 242: Please 1) discuss the purpose of the CFD simulations and 2) the uncertainties in

the SOLIDWORKS Flow simulations.

1) Purpose of the CFD simulations:

We have added a more detailed explanation of the questions relevant to adsorbent cartridge

sampling that we aimed to address with the simulations. The explanation is included in the

Discussion (Line 282), rather than the Experimental (Line 242, cited by the reviewer) where it

immediately precedes and contextualizes the results. The additional text reads as follows:

“The possible effects of air circulation created by the UAV multicopter rotors on the sampling

was considered. The flow field is also a factor in determining the sampler placement. There were

two main questions to be addressed. The first was to determine the time scale at which the air in

the sampling region beneath the UAV is flushed. If the flushing time scale is significantly less

than the sampling time, then, rather than being drawn from a stagnant pool, the sampled air can

be taken as representative of the surrounding air. The second was to determine the spatial scale

of the disturbance created by the rotors, in order to assess whether smoothing of concentration

gradients by rotor-induced mixing is likely to influence the measured values. Unlike many real-

time sensors, which have integration times on the order of a second, cartridge samples were

collected over relatively long time periods (minutes). Over this time period, atmospheric mixing

serves to average out gas concentration gradients at fine spatial scales (< a few m). Gradients at

this scale would therefore not be resolved by cartridge samples, even when not collected from a

UAV platform. If the spatial scale of mixing induced by the UAV is smaller than that of the

atmosphere itself over the sampling period, the perturbation of fine spatial scale gradients by the

UAV circulation will not significantly affect the measured concentrations. Hence, the second

critical question to be addressed by the CFD simulations is whether the spatial scale of

atmospheric mixing induced by the UAV rotors is larger than the spatial scale of atmospheric

mixing over the sampling period. If it is not, then the mixing due to the UAV should have little

effect on the cartridge samples.”

2) Uncertainties in SolidWorks Flow simulations

Some possible contributions to the uncertainty of the flow simulations are the domain size, the

grid spacing, the use of solid disks to simulate the rotors, and the landing gear position (down

instead of retracted). The domain size of +/-1 m and grid resolution were chosen to capture the

majority of the flow disturbance around the drone while also working within computational

limitations. For the same reason, sensitivity studies of the effect of changing the domain size or

grid spacing were not performed, so the uncertainties associated with variations in these

parameters are unquantified.

The magnitudes of the pressure variations around the drone (+/-100 Pa, or +/- 0.10%) speed

variations of ca. +/-0.2 m s-1 or ca. 2 to 25% of speeds of 1 to 12 m s-1. A 25% increase of the

calculated speeds would suggest a similar increase in the spatial scale for the dissipation of the

resulting disturbance. Hence, we estimate a range for the mixing scale of +/-5 to 7 m.

Other studies are consistent with the results of our simulations. Villa et al. (2016b) measured the

velocity fields around a smaller (3.7 kg) hexacopter and found that the downwash largely

dissipated within 3 m of the drone. Ventura Diaz and Yoon (2018) performed high resolution

CFD simulations of several quadcopter UAVs. The resulting velocity fields (cf. their Figure 10)

were qualitatively similar to those obtained in the current study, though the extent of the

perturbations was only +/-1 m. Both studies investigated smaller UAVs than used here. A larger

drone would be expected to have a larger mixing volume, consistent with the results of our

simulations.

Overall, allowance for possible uncertainties does not change the conclusion that mixing due to

the drone is likely less important than atmospheric mixing over the time period of the samples.

The following changes have been made to the text:

Section 2.5: “CFD simulations are carried out using SOLIDWORKS Flow Simulation (Ver. 2017

SP3.0) (Waltham, USA). Dimensions and an input geometric model of the UAV are obtained

from the DJI company (DJI Downloads). A box with the dimensions and location of the sampler

is added to the geometry file. The propellers are simulated by discs of the same diameter, and to

simulate a hovering UAV a downward velocity of 11 m s-1 is imposed through each disc so that

the lift produced by the motors balanced the system weight. The domain size was 2.4 m in width

and 2.0 m in height, with the UAV centered horizontally and at 1.2 m vertically. An adaptive grid

was used, such that the grid spacing is smaller where gradients are larger. Boundary conditions

include atmospheric pressure far from the UAV, which is set to 1 atm. As the actual pressure

during sampling may differ from this value, it is used only as a baseline for comparison. The

results are optimized by performing iterations until the pressure difference between the last two

iterations was within 2 Pa. Uncertainties in the CFD simulations could arise from the choice of

domain size or grid resolution, which were limited by available computational resources, or

assumptions such as the use of solid disks to model the rotors. In flight the legs are retracted to

horizontal. The simulations do not account for possible changes to the circulation patterns due

to the retraction of the landing gear, although this effect is expected to be minor minor relative

to the volume of the disturbance created by the drone (cf., Section 3).”

Section 3 (Results and Discussion): “The magnitudes of the pressure variations around the UAV

(+/-100 Pa, or +/- 0.10%) correspond to speed variations of ca. +/-0.2 m s-1 or ca. 2 to 25% of

speeds of 1 to 12 m s-1. A 25% increase of the calculated speeds would suggest a similar increase

in the spatial scale for the dissipation of the resulting disturbance. Hence, we estimate a range

for the mixing scale of +/-5 to 7 m.”

Line 264: It would be worthwhile to discuss the influence of rotor-wash potentially on

measurements and their differences at altitudes of 60 m, 75 m, and 100 m. Are these

measurements representative of 60 +/- 5m?

Here we address the question of the volume sampled by the drone as well as General Comment

1, above, which asks us to address “The dilution due to rotor-wash, which is a problem for all

instruments without an inlet that extends beyond the turbulence induced by the multi-rotor

platform.”

We agree with the reviewer that the volume represented by the measurement and the effect of the

UAV on this volume is critical to interpretation of the results. In contrast to previous studies, this

study does not aim to measure concentrations in a high-concentration plume emitted from a point

source into low-concentration background air with fast time resolution. Instead, we aim to

measure the average concentration from a horizontally varying non-point source over an

integration time of several minutes. We therefore think of the effect of the drone circulation as

‘mixing’ of concentration gradients in the surrounding air, rather than ‘dilution’, which suggests

loss of signal due to the introduction of background air into the sample. That is, there is spatial

averaging of the air sample within the mixing volume of the drone, but the sample itself is also

an average over the sampling time. The key question, as outlined in the ‘Purpose of the CFD

Simulations’ above, is whether the mixing volume due to the drone is larger or smaller than the

spatial scale due to atmospheric mixing of the air sampled over a 10 minute period. The revised

discussion of the drone mixing volume in the manuscript is included below. We conclude that

the mixing volume extends approximately +/-5 to 7 m above and below the UAV but that this

volume is small compared to the vertical scale of atmospheric mixing over the sampling time

period. Please also see the responses to Reviewer 2 regarding bias in the sample altitude and the

comparison of samples at different altitudes.

“Figure 4b shows the calculated air velocity distribution around the UAV. The simulation

suggests that air enters the sampling region experiences roughly laminar downward flow from

above the propellers, undergoes turbulent recirculation to the UAV sampling region, and then is

ejected below the UAV. The simulation shows that the air flushing time in the sample region is

fast (i.e., several seconds) compared to the timescale of VOC sampling (i.e., 5-10 min). The

velocity disturbance due to the rotors extends approximately 5 m above and below the UAV. This

is consistent with the CFD study by Ventura Diaz and Yoon (2018), which suggested that for

their smaller quadcopter (1.2 kg), the sample represented an air parcel extending approximately

1 m above the UAV. As expected for a larger drone, the disturbed air volume derived from

Figure 4 is significantly larger than in their study. The flow patterns are remarkably similar

considering the simplifying assumptions and lower grid resolution used in this study (cf. Section

2.5), lending credence to the general flow features shown in Figure 4. In addition, the simulation

shows that the air flushing time in the sample region is fast (i.e., several seconds) compared to

the timescale of VOC sampling (i.e., 5-10 min). The magnitudes of the pressure variations

around the UAV (+/-100 Pa, or +/- 0.10%) correspond to velocity variations of ca. +/-0.2 m/s or

ca. 2 to 25% of velocities of 1 to 12 m/s. A 25% increase of the calculated velocities would

suggest a similar increase in the spatial scale for the dissipation of the resulting disturbance.

Hence, we estimate a range for the mixing scale of +/-5 to 7 m. The simulations thus indicate

that the sampler performs representative real-time sampling of ambient VOC concentrations

averaged across several ±5 to 7 meters around the UAV. For comparison, the spatial scale of

atmospheric vertical mixing over the sampling period (10 min) can be estimated from the

relationship 𝑧 = √2𝐾𝜏, where K is the eddy diffusivity, τ is the time period, and z is the vertical

distance. Estimates of the eddy diffusivity within 10 m above a forest canopy are in the range of

approximately 2 to 15 m2 s-1 during the day, though the values are uncertain and vary with local

meteorology and canopy roughness (Bryan et al., 2012;Saylor, 2013;Freire et al., 2017). K then

generally increases with altitude for several hundred meters above the canopy (Wyngaard and

Brost, 1984;Saylor, 2013). Using the canopy-top values as a lower limit on the eddy diffusivity at

the UAV height results in an estimated lower limit on the vertical mixing scale of ca. 50 to 150

m, substantially larger than that due to the UAV. A manuscript treating atmospheric mixing

above the forest canopy more explicitly using a large eddy simulation (LES) method is currently

underway. Nevertheless, this estimate suggests that mixing due to the UAV is expected to exert

minimal influence on the measured VOC mixing ratios.”

Also note if these samples were taken on ascending vertical profiles or separate flights

(related to general comment 1).

The samples were collected on separate flights, as was stated in the original text (line 262):

“Three samples were collected in separate flights at heights of 60 m, 75 m, and 100 m relative to

the ground level at the tower location.”

Line 267: Were cartridges at the tower collected using an identical cartridge sampling

system, including a pressure sensor in the flow path and a mass flow sensor or only a

pump? Please describe this in the text.

No, the tower samples were collected using a hand-held motorized pump (Model 210-1002,

SKC). As this is a constant volume pump, pressure and temperature are needed to calculate the

total moles of sample air. For the tower samples reported here, temperature and pressure were

not measured simultaneously. A temperature of 25 C and pressure of 1.0 atm were used in the

calculation. Uncertainties in the temperature of +/-5 C (+/-2%) and pressure of +/-10% were used

to estimate the uncertainty in the mixing ratios. When combined with the other uncertainties, this

gives an overall uncertainty of 23% in the tower measurements. Table 1 has been updated with

the corrected mixing ratio values and uncertainties. The following changes to the text have also

been made:

“For comparison, VOC collections were performed concurrently atop the MUSA Tower with a

hand-held motorized pump (Model 210-1002, SKC). These samples were collected using a

volumetric flow rate of 200 sccm cm3 min-1 and sampling time of 20 min for a total sample

volume of 2.0 L. Mixing ratios were calculated from Eq. 1 using a pressure of 1.00 atm and

temperature of 25 ˚C (measurements of temperature and pressure were unavailable).

Uncertainties in pressure of +/-10% and temperature of +/-5 C (+/-2%) were used to estimate

an overall uncertainty of 23% for the tower samples.”

Line 285-290: Discuss in the text more explicitly what the impact is of deviations in

pressure in the sampling region. How would this specifically impact the representativeness

of cartridge measurements?

As discussed earlier, the mass flow sensor inherently accounts for changes in sample pressure

and temperature. Therefore, small deviations in the pressure of the sampling region should not

affect the measured total mass of air sampled, the resulting VOC mixing ratio, or the

representativeness of the measurements. To make this point in the text, the following sentence

has been added at line 285:

“Because the mass flow sensor inherently accounts for changes in sample pressure and

temperature, small deviations in the pressure of the sampling region should not affect the

measured total mass of air sampled or the resulting VOC mixing ratio. This result also suggests

that any possible effects of UAV pressure fields on any pressure sensitive sensor mounted in this

area would be small.”

L346-347: This second half of this sentence is a bit confusing. Isn’t pre-programed GPS-

based operation already employed? Is the goal to integrate that seamlessly into the DJI

flight software?

The long-term goal is to control the sampler from the remote controller or flight-control app

through the drone’s signal output. For the first generation sampler described in the manuscript,

however, the drone flight was controlled by pre-programmed GPS-based operation, but there was

no communication between the drone or remote controller and the sampling box. The sampler

was programmed to open and close the sample valves at pre-determined times after takeoff. The

GPS control program was synchronized with it based on these elapsed times (with an added

buffer). For example, if the drone flight time to the first sampling point was 2 minutes from

takeoff, the first sample would be initiated 3 minutes into the flight and last for 10 minutes. To

clarify this point in the manuscript, we have revised the text as follows:

“A major goal of ongoing development of the sampler is to enable operation control of sampler

functions and collection of sampler data from the tablet-based drone control software, either

manually or as part of a pre-programmed GPS-based flight trajectory algorithm. In the current

version, the flight trajectory is programmed with the drone control software, whereas and

sampler operation is controlled by a stand-alone program on the Arduino Uno microcontroller.

The two programs are synchronized in time from initialization with a short time buffer so that

the drone arrives at the sampling location 1 min prior to opening the valve. Both of these

operational modes require In order to fully integrate these functions, real-time communication

among the sampler, the UAV on-board computer, and the user control interface on the tablet is

required. The Arduino Uno microcontroller is unable does not have the capability to

communicate with the UAV on-board computer. To address this issue, an ongoing the next step

in the development is the replacement the Arduino Uno microcontroller with a Raspberry Pi

miniature computer.”

L356: How high were the winds on these days that operation of the solenoid, pump or

sensors failed? How typical are winds this high?

On the days the sampler failed, the wind speeds were around 5 m/s. Winds > 4 m/s for short

periods are observed relatively frequently (40-50% of sampling days). The sampler, however,

does not always fail under these conditions. The failure rate over 128 flights (including flights

after the period reported in the manuscript) is about 2.5%. In addition, changes made to

ruggedize electrical connections in the sampling box and frequent inspection of the electrical

connections (before each flight) have largely addressed this issue. To incorporate these points,

we have amended the text as follows:

“This capability can be important to alert the user to problems during flight, such as the failure

of valves or the pump to be activated, as has occurred occasionally on windy days (5% of flights

with winds >4 m/s) due to strong vibration. This failure mode has largely been eliminated by

reinforcing the electrical connections and inspecting them before each flight.”

P22 (Figure 4): The M600 Pro is not typically flown (and I imagine samples aren’t

collected) with the legs down for landing. How is the flow in these simulations altered when

the M600 legs are retracted, if at all? See general comment 1.

All samples were collected with the landing gear retracted. The reviewer raises a good point that

the circulation patterns around the drone could be somewhat different with the landing gear

retracted than in landing position. The simulations were run with the landing gear down because

they are in that position in the CAD files provided by the manufacturer. Unfortunately, for

logistical reasons, it would be difficult to run new simulations with the landing gear retracted.

The co-author who ran the simulations (J. Baptiste) is no longer at Harvard, where the original

simulations were run. In addition, for licensing reasons, we no longer have access to the software

package that was used previously.

Hence, we will address the reviewer’s concern using alternate approaches.

First, as was discussed above, the key question in the context of cartridge sampling is whether

the drone creates atmospheric mixing on a spatial scale larger than the atmospheric mixing that

takes place within the sampling period. The conclusion is that the spatial scale of the air sampled

over a 10 minute period due to atmospheric mixing is larger than the ca. +/-5 m mixing scale of

the drone. The position of the landing gear in the simulations becomes an issue if it changes the

mixing scale enough to change the answer to this question.

The landing gear are composed of slender carbon fiber rods. As Figure 4 shows, air is drawn

downward from above the drone through the rotors. It then recirculates upward in the region

beneath the drone where the sampler is mounted. Based on the figure, the absence of pressure or

velocity gradients in the immediate vicinity of the legs suggests that the presence of the legs does

not significantly perturb this flow. We therefore conclude that the position of the landing gear is

unlikely to significantly alter the mixing scale suggested by the simulations.

Further, we have added a reference to a paper by Villa et al., (2016b), who measured the velocity

fields around a smaller hexacopter drone (3.7 kg vs. 9.6 kg + 1.0 kg payload in the current

study). The velocity fields deduced from their measurements show overall flow patterns

consistent with the simulation results shown in Figure 4, although the spatial scale of the

disturbance would be larger for a larger and heavier drone.

We have added the following sentence acknowledging the possible effect of the landing gear

position to the manuscript (Line 250):

“In flight the legs are retracted to horizontal. The shown simulations do not account for possible

changes to the circulation patterns due to the retraction of the landing gear, although this effect

is expect to be minor relative to the volume of the disturbance created by the drone (cf. Section

3).”

P22 (Figure 4): Please add a vertical scale and horizontal scale on Fig. 4a and Fig. 4b.

The figure has been revised to include vertical and horizontal scales. We have also added a

figure caption (below), which was inadvertently omitted in the earlier version.

“Figure 4. (a) Vertical pressure distribution and (b) air velocity distribution around the UAV

from the CFD simulation. Pressure difference between the UAV sampling area and the area

under the propellers was simulated to be less than 100 Pa indicating a minimal effect of pressure

on sampling. The air velocity was 1.65 m s-1 upward around UAV sampling region, suggesting a

fast air flushing time underneath the sampling box.”

Reviewer 2:

General Comments: This is a very well-written manuscript describing the development of a

VOC sampler for autonomous, drone-based sampling. The motivation and relevant

background is thoroughly but succinctly presented in the introduction, and the platform

and results are clearly and generally well-described. I recommend publication of the

manuscript, pending the authors: 1) add some context for what results should be expected

for vertical distribution of VOCs in Table 1, so that the reader can better interpret the

results presented here, and 2) more satisfactorily explore the vertical sampling bias

introduced by rotors drawing air down from above (or gather comments from an

additional reviewer with substantial experience with the fluid dynamics of drones). The

CFD analysis is laudable, but does not conform to experience in working with large drones

C1 with payloads, where vertical disruption of plumes extends greater than 5 m in many

cases, and the paper cited to suggest < 1 m disruption is based on drone platforms that are

substantially smaller.

We thank the reviewer for the thoughtful and helpful comments, which have led to substantial

improvements to the manuscript. Responses to individual comments, including the two in the

summary paragraph above, and the corresponding manuscript revisions are detailed below.

Specific comments:

111 – Noteworthy that the sampler was placed on the platform underneath the drone.

Downwash and eddies present a significant challenge in sampling underneath drones (as

you explore later), leading many to mount sensors on top of the drone, where flow is

laminar, or to extend a sampling inlet outside the rotor influence. CFD simulations are a

helpful place to start, but ultimately you can learn a lot by just flying your specific

platform through a smoke plume. You’ll notice straight, laminar flow lines on top that

extend from several meters above (depending on system mass) and a mess of eddies

underneath. Dave Barrett and Scott Hersey at Olin College of Engineering presented on

this in collaboration with Aerodyne at AAAR and AGU in 2016 – check their materials for

more clues. This eddy issue matters less for your application than for their 1-Hz

instrument, since you are not after time-dependent (i.e. highly spatially resolved) data, but

rather bulk VOC mass over an entire flight segment. But is nonetheless an important

consideration. Explore options to mount on top, or to extend a sampling inlet to a point

horizontally outside rotor influence.

We agree with the reviewer that there are potential drawbacks to mounting the sampler beneath

the drone. There are also advantages. Likewise, there are advantages and disadvantages to

mounting it on top. One particular disadvantage to top mounting is that we have observed that

the temperatures at the top surface of the drone can get extremely hot, particularly during the dry

season. This could have a particularly detrimental effect on adsorbent cartridges due to the

higher volatility of VOCs at higher temperatures. As noted by the reviewer, the presence of

eddies underneath the drone is less of an issue for our application, where samples are collected

over a 10 minute period. After weighing these factors, we conclude that the choice to mount the

sampler beneath the drone is a reasonable one for this particular application. We have added a

discussion of these issues to Section 3 of the text, as quoted below:

“There are both advantages and disadvantages to mounting the sampler either atop or beneath

the UAV. The advantages of top mounting include faster time response and potentially higher

spatial resolution due to laminar flow and less mixing. Some disadvantages are the potential for

bias in some measurements, such as of particles, due to sampling from laminar flow rather than

well mixed air, and the potential for more vertical bias due to the strong laminar downwash of

air above the UAV. In addition, the temperatures at the top surface of the UAV have been

observed to become extremely hot (ca. 40 ˚C), particularly during the dry season. This is

particularly problematic for collecting VOCs on adsorbent cartridges, as the sampling efficiency

may be reduced at elevated temperatures. On the other hand, the advantages to mounting

beneath the UAV are that the sampler is protected from direct sunlight and therefore cooler.

Also, the flow beneath the UAV is well mixed, which avoids flow effects such as a bias towards

large particles. Disadvantages, such as mixing of concentration gradients and decreased time

resolution, are most significant for sensors with fast time response. A study by Villa et al.

(2016b), however, explored the differences in measured concentrations of a suite of trace gases

from a point source when the sensors were mounted above, below, and in the horizontal plane of

a hexacopter UAV. Their results show similar dilution of the plume measured above and below

the UAV, suggesting that the air sampled on top of the drone does not necessarily experience

less mixing. A sample inlet mounted such that it extends horizontally outside of the rotor wash

was the least affected by the UAV flow fields and could be a good solution for fast sensors. The

presence of eddies underneath the drone is less of an issue for our application, where samples

are collected over a 10 minute period. Atmospheric mixing and temporal averaging will smooth

out mixing ratio gradients over this time period, so mixing by drone-induced eddies should have

little effect on the measurement. Since the disadvantage of overheating if the sampler is mounted

on top of the UAV potentially outweighs the disadvantage of sampling from the turbulent flow

underneath, the decision to mount the sampler beneath the UAV is a reasonable one for this

particular application.”

240 – CFD simulation parameters are described, though it’s not explicit at this point why

you did CFD simulation (I can assume where you’re headed). I suggest giving some sense of

the need/purpose for this simulation before introducing it.

We have added material to the introduction to discuss the need for the CFD simulations to

understand the flow fields around the drone and their possible effects on the measurements and

in the discussion to describe the specific aims of the simulations. These changes are described in

more detail in the response to Reviewer 1, who made a similar comment.

258 – The drone was launched and recovered from a platform above the canopy, but one of

the key motivations for the drone-based sampling platform is to avoid the need for

platforms and to be able to access more remote sampling locations. Can you speak to the

usability of this platform in the types of contexts that motivate the study (i.e. those with

dense canopies and no platforms)?

There were several reasons for launching the drone from the tower in this study. The first was

inexperience. Until we gained expertise in flying the drone with the sampler, we were most

comfortable maintaining visual contact with the drone. Secondly, in many places (including the

US) regulations for the use of UAVs require that the pilot maintain visual contact. This may

change in the future as the use of drones becomes more widespread. For flights without visual

contact, a camera would be useful for visualizing the position of the drone. In order to reduce the

payload weight, no camera was mounted on the drone during sampling flights. This could be

changed by adding a small camera at the expense of a few minutes of flight time or by using a

second drone with a camera. In order to fly in an area with a dense canopy and no tower, it

would be necessary to have at least a small clearing in which to take off and maneuver the drone

up through the canopy. With additional experience and a camera for visualization, this should be

possible in the future.

The following text has been added to the discussion:

“Current regulations in some locations, including the US, require that the operator maintain

visual contact with the UAV. This was also deemed best practice in the current study as users

gained experience and comfort with flight operations. Launching the UAV from a tower

permitted the pilot to maintain visual contact during flight. As another approach, the UAV

sampler has also been flown in locations with hills where it is possible to visualize the top of the

canopy over an area of lower elevation from an area of higher elevation. In the future, as

regulations permit, navigation from the ground to above the canopy should be possible and

would allow sampling in more remote and densely forested regions. A clearing of sufficient size

to allow the UAV to be navigated would be required. A camera to provide remote visualization,

either on the same drone or on a second companion drone, would aid in navigation outside of

the pilots visual range.”

262 – Given the note above, and the high velocity of air flow down through the rotors of the

drone, I am not convinced that 60 m actually represented 60 m. I should be clear that I see

your exploration of this with CFD modeling, but your model results conflict with my

experience seeing drones sample smoke plumes in the field. With a slightly larger drone

(S900) and slightly heavier payload (2.5 kg), I consistently see rotors draw down air from

several (>/= 5) meters above mounted instruments in buoyant plumes. Experience suggests

to me that your vertical sampling bias is greater than the 1 m suggested in line 294.

Further, the result suggesting 1 m vertical bias in air sampling based on rotor air flow in

Diaz and Yoon (2018) is based on a significantly smaller drone with no payload. Your large

drone with payload will, necessarily, exert a greater vertical impact on air flows than

theirs. This comment comes with the caveat that I am basing them solely on experience and

observations with quad copters, and no modeling or detailed analysis of my own. I

recommend either a brief review of this section – especially as it relates to altitude-of-

sample bias – by a reviewer with greater expertise in the fluid mechanics of multi-rotor

aircraft, or an addition of language that outlines the potential for vertical sampling bias on

the order of several meters.

We agree with the reviewer, both that the vertical mixing volume is larger than +/-1 m around

the drone, and that there is likely a bias in the sampling height due to the downward motion of air

induced by the drone. The perturbation volume question is addressed in more detail in the

response to Reviewer 1’s comment on Line 262. To address the question of vertical bias, we

have added the following text to the Discussion:

“As noted above, the sampled air is drawn systematically from above the altitude of the UAV. It

is therefore expected that the sampled air represents an altitude slightly higher than the flight

altitude. Based on a mixing volume extending 5 - 7 m above the drone, a vertical bias of ca. -3 m

altitude is inferred.”

278 – “Reasonable consistency” is subjective. Quantify, and compare with either

sampling+measurement uncertainties or previously published variability in VOC

concentrations with height above canopy (or both).

We have replaced the sentence referenced by the reviewer with the following:

“Nevertheless, the results demonstrate reasonable consistency between samples collected by the

UAV and on the tower, separated by 711 m. They also suggest that vertical concentration

gradients can be assessed using this method. The results for all fall within the expected range of

concentrations (e.g., ca. <1 – 10 ppb for isoprene) for the near-canopy environment over the

Amazon rainforest based on previous observations (Alves et al., 2016; Harley et al., 2004).”

282 – CFD modeling appears. I applaud the authors for attempting to address rotor

influence in sampling. Ultimately, as I stated above, I expect the below-drone air flow

perturbations to be less important for your application of 10 min resolution samples. But

the bias introduced in the vertical resolution is of concern and my experience tells me that

for a drone your size, the vertical extent of air disruption is substantially greater than the 1

m suggested here, based on results from a much smaller drone platform with no payload. I

am, unfortunately, not the right reviewer to critique your CFD model run, and suggest that

an additional reviewer explore this.

We thank the reviewer for the helpful comment and agree with all points. The question of

vertical bias in the sampling height is addressed in response to the previous comment by this

reviewer on Line 262. The vertical extent of air disruption is discussed in response to Reviewer

1’s comment on Line 262.

Table 1 – Can you put these results in context that help the reader understand the

consistency of measurements and how they conform to expectation? For example, I notice

that isoprene concentrations vary substantially with altitude, though not in a way that

decays with altitude (as I might expect). Same with Pinene(s). As presented, I’m unable to

discern why the 100 m sample at the sampling site has higher concentrations of

monoterpenes than both the 60 m and 75 m sample. Can anything be determined from

ratios of VOCs to tell what’s going on here? What should I expect to see in vertical

variability? This doesn’t conform to my expectations of reducing concentration with

altitude, so please explore this so that the reader isn’t left with questions about whether

sampling bias or the drone platform is responsible.

For samples collected simultaneously at different altitudes above a single location, we would

indeed expect a gradient of decreasing concentrations with height. Other variables can, however,

influence concentrations in different locations, such as different canopy sub-types with different

emission rates. VOC emissions also respond strongly to changes in light and temperature, so

concentrations at a single location can vary strongly over periods of a few hours or even minutes.

As a result, it is difficult to make direct comparisons between the samples presented in Table 1,

which were all collected at different locations (tower vs. point A), altitudes, and times. For

example, as mentioned, the 100 m sample at point A has a higher concentration than those

collected at 60 and 75 m, but it was also collected closer to early afternoon (13:15 – 13:35 h),

when VOC emissions typically peak, than were the 75 m (11:15 – 11:35 h) or 60 m (15:15 –

15:35 h) samples.

More samples with systematic vertical, horizontal, and temporal coverage and a modeling

framework incorporating emissions, atmospheric mixing, and chemistry are needed in order to

draw firm scientific conclusions about the implications of atmospheric variability across these

coordinates. Such sampling and analysis is currently underway and the results will be explored

further in subsequent publications.

To address this question, we have added the following explanation to the text:

The results for isoprene all fall within the expected range of concentrations (ca. <1 – 10 ppb) for

the near-canopy environment over the Amazon rainforest based on previous observations (Alves

et al., 2016; Harley et al., 2004). VOC emissions concentrations depend on many conditions,

including season, time of day, temperature, light levels (i.e., cloudiness), height above the

canopy, and canopy forest composition, which can vary on spatial scales of 10’s of meters.

Nevertheless, the results demonstrate reasonable consistency between samples collected by the

UAV and on the tower, separated by 711 m.They also suggest that vertical concentration

gradients can be assessed using this method. Atmospheric concentrations are also affected by

atmospheric turbulent mixing and photochemistry. It is therefore difficult to make direct

comparisons among the samples presented in Table 1, which were all collected at different

locations (tower vs. point A), altitudes, and times. More samples with systematic vertical,

horizontal, and temporal coverage and a modeling framework incorporating emissions,

atmospheric mixing, and chemistry are needed in order to draw firm scientific conclusions about

the implications of atmospheric variability across these coordinates. Further analysis and

scientific interpretation of these results and a larger data set are the subject of separate

forthcoming publications.

References

Alves, E. G., Jardine, K., Tota, J., Jardine, A., Yãnez-Serrano, A. M., Karl, T., Tavares, J.,

Nelson, B., Gu, D., Stavrakou, T., Martin, S., Artaxo, P., Manzi, A., and Guenther, A.:

Seasonality of isoprenoid emissions from a primary rainforest in central Amazonia, Atmos.

Chem. Phys., 16, 3903-3925, 10.5194/acp-16-3903-2016, 2016.

Bryan, A. M., Bertman, S. B., Carroll, M. A., Dusanter, S., Edwards, G. D., Forkel, R., Griffith,

S., Guenther, A. B., Hansen, R. F., Helmig, D., Jobson, B. T., Keutsch, F. N., Lefer, B. L.,

Pressley, S. N., Shepson, P. B., Stevens, P. S., and Steiner, A. L.: In-canopy gas-phase chemistry

during CABINEX 2009: sensitivity of a 1-D canopy model to vertical mixing and isoprene

chemistry, Atmos. Chem. Phys., 12, 8829-8849, 10.5194/acp-12-8829-2012, 2012.

Freire, L. S., Gerken, T., Ruiz-Plancarte, J., Wei, D., Fuentes, J. D., Katul, G. G., Dias, N. L.,

Acevedo, O. C., and Chamecki, M.: Turbulent mixing and removal of ozone within an Amazon

rainforest canopy, Journal of Geophysical Research: Atmospheres, 122, 2791-2811,

doi:10.1002/2016JD026009, 2017.

Goldstein, A. H., and Galbally, I. E.: Known and unexplored organic constituents in the earth's

atmosphere, Environmental Science & Technology, 41, 1514-1521, 10.1021/es072476p, 2007.

Harley, P., Vasconcellos, P., Vierling, L., Pinheiro, C. C. d. S., Greenberg, J., Guenther, A.,

Klinger, L., Almeida, S. S., Neill, D., Baker, T., Phillips, O., and Malhi, Y.: Variation in

potential for isoprene emissions among Neotropical forest sites, Global Change Biology, 10,

630-650, 10.1111/j.1529-8817.2003.00760.x, 2004.

Pankow, J. F., Luo, W., Melnychenko, A. N., Barsanti, K. C., Isabelle, L. M., Chen, C.,

Guenther, A. B., and Rosenstiel, T. N.: Volatilizable Biogenic Organic Compounds (VBOCs)

with two dimensional Gas Chromatography-Time of Flight Mass Spectrometry (GC × GC-

TOFMS): sampling methods, VBOC complexity, and chromatographic retention data, Atmos.

Meas. Tech., 5, 345-361, 10.5194/amt-5-345-2012, 2012.

Saylor, R. D.: The Atmospheric Chemistry and Canopy Exchange Simulation System

(ACCESS): model description and application to a temperate deciduous forest canopy, Atmos.

Chem. Phys., 13, 693-715, 10.5194/acp-13-693-2013, 2013.

Ventura Diaz, P., and Yoon, S.: High-Fidelity Computational Aerodynamics of Multi-Rotor

Unmanned Aerial Vehicles, in: 2018 AIAA Aerospace Sciences Meeting, AIAA SciTech Forum,

American Institute of Aeronautics and Astronautics, 2018.

Villa, T. F., Salimi, F., Morton, K., Morawska, L., and Gonzalez, F.: Development and

Validation of a UAV Based System for Air Pollution Measurements, Sensors, 16, 2202, 2016b.

Woolfenden, E.: Sorbent-based sampling methods for volatile and semi-volatile organic

compounds in air: Part 1: Sorbent-based air monitoring options, Journal of Chromatography A,

1217, 2674--2684, 10.1016/j.chroma.2009.12.042, 2010a.

Woolfenden, E.: Sorbent-based sampling methods for volatile and semi-volatile organic

compounds in air. Part 2. Sorbent selection and other aspects of optimizing air monitoring

methods, Journal of Chromatography A, 1217, 2685--2694, 10.1016/j.chroma.2010.01.015,

2010b.

Wyngaard, J. C., and Brost, R. A.: Top-Down and Bottom-Up Diffusion of a Scalar in the

Convective Boundary Layer, Journal of the Atmospheric Sciences, 41, 102-112, 10.1175/1520-

0469(1984)041<0102:Tdabud>2.0.Co;2, 1984.

Zhao, Y., Kreisberg, N. M., Worton, D. R., Teng, A. P., Hering, S. V., and Goldstein, A. H.:

Development of an In Situ Thermal Desorption Gas Chromatography Instrument for Quantifying

Atmospheric Semi-Volatile Organic Compounds, Aerosol Science and Technology, 47, 258-266,

10.1080/02786826.2012.747673, 2013.